[ Saturday, January 22, 2000 ] Jon Preston wrote:
> My config is:
> 
>     2x PIII 500
>     128MB Ram
>     DPT PM3754U2 Raid V Millennium
>     3 - IBM 9.1GB Ultra2 LVD 80MB/s
>     Red Hat 6.1
> 
>    "request_module[block-major-9]: Root fs not mounted

Note that block-major-9 is for /dev/md*, the *software* raid
devices.  This would seem to indicate that it's trying to
mount root as /dev/md0 or something similar.  Perhaps this
was mis-typed-in?

>     VFS: Cannot open root device 08:02
>     Kernel panic: VFS: Unable to mount root fs on 08:02"

08:02 is /dev/sda2 which makes more sense.  What I'd imagine
is happening is that the support for your RAID controller
was built as a module and that module's not loaded when it's
trying to mount root.  Two possible reasons:

1) initrd was used but that module was mistakenly left out.
   mkinitrd will automatically include any modules aliased
   to scsi_hostadapter so you may need to check conf.modules
   and possibly add something like "alias scsi_hostadapter dpt"
   replacing dpt with the module name for your controller
   and rebuilding your initrd with mkinitrd and re-running lilo

2) Not using initrd and support for the controller was mistakenly 
   left out of the kernel configuration.  I have to admit I don't
   see anything in my kernel configuration that looks close, so I'm
   not sure where in the config to check for this.

>     /dev/sda1    (mount point)= "/boot"             15M  Type 83
>     /dev/sda2    (mount point)= "/"                4000M  Type 83
>     /dev/sda3    (mount point)= "/data"        13654M  Type 83
>     /dev/sda4    (Swap)                           "remainder" Type 82

It looks to have found the kernel in /boot (unless you're booting off
of a floppy or similar)  Since the 6.1 install process ran, support
must exist for this controller (it's the same controller used during
install, right?)  so it should just be a matter of getting the kernel to
have the necessary device support when it's trying to mount root since
the VFS messages seem to indicate that it's looking in the right place.

> I've tried to boot with sda1 as the active partition and sda2 as the active
> partition.

sda1 should be it (not that it'll much matter) as your booting process
is running off MBR at the beginning of the drive and the kernel that's
in /boot on sda1.

Good luck,

James Manning
-- 
Miscellaneous Engineer --- IBM Netfinity Performance Development

Reply via email to